Measuring instabilities and chaos in the real world: applications in rotating machinery

Abstract

Rotor dynamic instabilities have resulted in seldom understood and costly failures in high-performance rotating machinery. At times, machines test perfectly in the lab, yet fail under load and cannot achieve their designed capacities in the field. Nonlinear dynamics offers a way to help understand and possibly prevent these problems. Computed order tracking prepares experimental data for dynamical analysis by resampling data in real time to match a time-varying forcing function which in this instance is the rotational speed of the machine. A commercially available instrument can be used to view Poincare diagrams of a rotor test set as it begins to exhibit signs of instability.<>